Europe’s Banking Union Has a Hidden Border
This series examines the incomplete European Banking Union through a practical question: how can Europe remain financially fragmented even when the loans that do cross borders carry similar prices? This opening focuses on the distinction between access to a banking relationship and the terms of a loan after that relationship exists. Understanding that distinction gives the reader a better way to read claims about European financial integration: look first at who can operate, who can connect, and only then at the price of the contracts that survive those hurdles.
An open market on paper, a segmented market in practice
The European Union has long treated the free movement of capital as a basic principle. The euro removes exchange-rate risk among member countries that share it, and the Banking Union has placed important parts of supervision and resolution on a European footing. Those changes matter. They make cross-border finance more feasible than it would be if every bank had to navigate entirely separate monetary and supervisory systems.
The reforms also changed the question facing a bank. A group no longer needs to regard every national border as the edge of a wholly separate monetary world. It can, at least in principle, raise funds, extend credit, and organize operations across a more integrated area. But an area that permits a transaction is not necessarily an area in which the transaction is easy to originate, monitor, insure, resolve, or scale. Formal permission removes one class of barrier; it does not erase the practical work required to make a loan relationship viable.
That practical work is particularly important in banking because credit is not a standardized good delivered without context. A lender has to assess a firm's accounts and prospects, understand the legal environment for collateral and bankruptcy, monitor performance, handle reporting, and decide how much capital and liquidity to commit. A borrower has to find a lender willing to learn about its business and to remain accessible when conditions change. These activities are shaped by national environments. Banks can encounter different prudential requirements, resolution arrangements, deposit-insurance systems, supervisory practices, governance rules, entry conditions, accounting conventions, and bankruptcy regimes. A common currency does not make those operating conditions identical.
That gap is the subject of the IMF Working Paper Barriers to a European Banking Union. The paper does not treat “fragmentation” as a vague feeling that countries remain different. It studies concrete relationships: firms borrowing from banks, banking groups establishing branches or subsidiaries abroad, and the country-pair conditions that stand between them. The practical consequence is broad. If savings cannot reach firms through a wider set of intermediaries, the issue is not confined to the financial sector. It can affect the cost of financing capital, the scale of investment, labor demand, and the growth opportunities available in different parts of Europe.
A loan price is the end of a chain
The most important distinction is between the extensive and intensive margins of banking. The extensive margin asks whether an event happens at all: does a banking group enter a foreign market, or does a particular firm form a lending relationship with a foreign bank? The intensive margin asks what happens after the event: what interest rate is charged, how large is the loan, and how much of the firm’s borrowing comes from that bank?
These are different questions with different economic meanings. Imagine a firm in one euro-area country considering a bank headquartered or operating in another. Before anyone can compare interest rates, the bank must be willing to serve that market, the firm must know that the bank is available, both sides must be able to exchange information and satisfy their respective requirements, and the proposed relationship must survive the screening process. If the link is never formed, there is no observed loan rate to place in a table. The absence is economically meaningful even though it leaves no interest rate behind.
Now imagine comparing the interest rates on two sets of loans that have already been signed. If the rates are close, that may indicate that banks price the surviving relationships in similar ways. It does not tell us how many firms considered a foreign bank and failed to connect, how many foreign banks never entered the market, or how many potential loans disappeared before they could be priced. The comparison begins after several decisions have already filtered the market.
This is why price convergence can coexist with fragmentation. A foreign bank may offer competitive terms to a large, mature firm with which it has managed to establish a relationship. The observed loan can look no more expensive—and perhaps slightly cheaper—than a domestic loan. Yet smaller or less familiar firms may never reach that stage, and a bank that has not entered the country cannot compete for them in the first place. The market can therefore be integrated for the narrow group of realized cross-border links while remaining nationally organized for most potential borrowers and lenders.
The same logic applies on the bank side. A large banking group may be capable of lending abroad, but capability is not the same as a local operating presence. Entering through a branch or subsidiary commits the group to an institutional arrangement. It may have to understand host-country rules, coordinate with supervisors, account for deposit protection and resolution, and decide how much of its balance sheet can move across the group. If it chooses not to enter, firms in the destination country face a smaller set of possible lenders even if capital is legally free to move.
This distinction also guards against a common mistake in interpreting averages. A small conditional difference in interest rates is not evidence that financing opportunities are equally available. It describes the terms among borrowers and banks that have cleared the access hurdle. The relevant question for the Banking Union is wider: whether institutions and firms can form the links that allow those terms to be offered in the first place.
The access infrastructure behind the contract
A useful way to picture the subject is as a chain of dependencies. National regulatory and operational differences can affect whether a banking group establishes a foreign presence. That presence changes the set of lenders available to firms. The available set affects whether a firm and a bank form a relationship. Only then do the familiar contract outcomes—price and quantity—become observable. Those financing conditions can in turn affect the firm’s effective cost of capital and its decisions about investment and employment.
The order matters because each stage changes the opportunities at the next stage. A firm cannot choose among foreign banks that do not operate in its market. It cannot compare a foreign offer that was never made. And a bank cannot price a relationship that was blocked by an entry decision, a compliance cost, an information gap, or a failure to match with a suitable borrower. Looking only at realized contracts can therefore make the final stage appear more integrated than the chain that produced it.
This is not a claim that every missing link is caused by regulation. Geography, language, culture, information, borrower risk, bank capacity, and group strategy can also shape cross-border activity. Nor does it mean that every observed foreign loan is evidence of an efficient allocation of capital. It means that the Banking Union has an access infrastructure, and that infrastructure can fail before a price comparison becomes possible.
The distinction is useful for policy as well. If a problem lies mainly in the price of loans after matching, rate competition or funding conditions may be the relevant focus. If the problem lies in entry and relationship formation, reducing the cost of understanding and operating across national systems may matter more. Those reforms can have different beneficiaries: a pricing change helps firms already connected to lenders, while a change in access can alter which firms and banks enter the market at all.
The arrows in this map are dependencies, not a claim that one institution controls every outcome. European reforms can reduce some barriers while leaving others in place. National rules can raise costs without being the only reason a bank stays home. A relationship can form for reasons that are not captured by a country-pair average. The value of the map is that it keeps these stages separate long enough to ask which one the evidence actually observes.
The evidence in this series follows the same logic. It begins by establishing what counts as a cross-border relationship and what the observed banking network looks like. It then separates the probability of forming a relationship from the terms of an existing loan and from the decision of a banking group to enter. Next, it examines how differences in national banking rules line up with those barriers. Finally, it uses a calibrated model to ask how a reduction in access frictions could affect output—and whether the gains would come from better allocation, more capital and labor, or both.
Each step carries a different level of certainty. The banking network and loan relationships are observed in administrative data. The connection between regulatory distance and estimated barriers is an association, not a clean causal estimate of one reform. The economy-wide effects are predictions from a model whose parameters and structure are calibrated to the data. Keeping those categories separate is essential: otherwise a conditional rate comparison can be mistaken for a market-wide fact, or a model experiment for a forecast.
There is also a difference between saying that integration is valuable and saying that every border should disappear in the same way. Regulation can protect depositors, limit risk, and determine who bears losses when a bank fails. A more integrated market may bring broader access while also changing supervisory responsibilities and the distribution of risk. The question is therefore not whether national institutions have any legitimate role. It is whether persistent differences impose unnecessary barriers on otherwise viable relationships, and whether reducing those barriers produces benefits that reach beyond the already connected.
The central question, then, is not simply whether European loan prices have converged. It is whether European firms and banks can reach one another across the borders that remain embedded in operations, regulation, and organization. Before judging the terms of cross-border credit, we need to see the network that makes those terms possible—and the home bias that keeps much of it out of reach.